tsunami

log in
history

Thoughts on Google

Luke Breuer
2008-01-13 22:14 UTC

In a nut shell
Google makes most of its money on advertising [citation needed]. Therefore, it ma


I often run into limitations of Google's search engine. In particular, I have a penchant for using search terms which have another, more popular meaning on the web. Since Google primarily ranks by popularity (the more people with more authority who link to you, the higher your site will be ranked), this is a problem.

A very desirable feature is one that would tell me how my search results relate to each other; one way to do it is via clustering, like Vivísimo. This is one way to overcome the popularity issue mentioned above.

I often wonder whether Google has economic reasons to not improve its search engine quality any more than to keep ahead of competition. In particular, does it make money on advertisements due to users not being able to retrieve the most relevant search results?

I'm not convinced Google will be able to improve appreciably on search without powerful linguistics abilities, abilities I think will only come with strong AIs. Seeing as strong AIs will not be around in the next decade and probably not the next few decades, it seems humans will have to help categorize and organize information. At a lecture by a Google engineer, I heard that Google is trying to keep its algorithms as free from human influence as possible; this is a problem.

The ultimate capability would be something like:
select  distinct
        idea
from    Somewhere
where   topic = 'topic of interest'


Interestingly, Steve Yeggie thinks Google is approaching AI via search. There's a bit in there about the semantic web not working because people are lazy.